fix for issue #25, #26, #27 and warning cleanup
Fixed imports to avoid circular import issue #27
more fixes for create_new flag tests
fixing tests for create_new flag
DSS file only created when explicitly create_new=True(default: False)
bugfix for empty arrays
handle pandas Series on writes
using generators for get_ts and get_matching_ts functions for efficiency context manager for use "with" statement fixed for pandas >=1.1 auto versioning turned on
fixed bug when switching to non-condensed catalog that caused shortening of time windows
- fix for breaking change by pandas >= 1.1
- switching to using non-condensed catalog, flag available to switch back
- changing exception on missing pathname to logging message
- changing message level to 0 as default
- adding get_rts_matching for pathname selection by regular expressions
- replaced warnings with debug logging
Fixed for '*' in catalog file in case of missing periods
Added helper function for simple timeseries retrieval
regular and irregular timeseries now share same time window parsing logic
fixes for issue #16: pandas timestamp limitations
another fix for pandas 1.0.0 upgrade - issue #15
fix for issue 15: units 'M' and 'Y' no longer supported
Fixed issue with end of timestamp writing to dss files for "PER-*" data type
Recompiled heclib in linux with latest compilers to resolve issue 8
Partial fixes for offset: Fixed for INST-* timeseries but not PER-* timeseries (issue #12)
Performance tests added to showcase pyhecdss is the fastest Fixed issue #10: Period data stored shifted to end of time stamp (HEC-DSS convention) Fixing libgfortran dependency on linux by statically linking the library
Merged pull request from HenryDane: issues warnings or throws exception based on return value from functions fixed minor typo(s), added tests
Fixes issues #2, #4: get_version(), set_message_level(), set_program_name() now supported at module level Fixes issues #1: HECDSS marks periods by end of time stamp bug
Added linux 64bit support Corrected license to MIT for conda distribution
Update tests to use a smaller test.dss files Added sphinx documentation and demo notebook
Write irregular time series
Write regular time series Read irregular time series as data frame + units + type Performance improvement to using np.zeros instead of np.array(range...)